goessnd [ parameter=value ... ] [ outputfile ... ] goessnd [ parameter=value ... ] [ directory ]
Parameters are: line_sub, samp_sub, average, max_pass, climatology, weight, debug, skin_type, min_clear, gamma, wv_scale, guess_file, ATEST, BTDIFF, DF, EMISS, EMLIM, EROLIM, GCK, TINC, TSKLIM, ULIM, VIS, WFAC, WLIMX, ADJ, CGES, GAMS, ISAVG, ISUN, IVV, NOBL, NOSFC, NOSK, WOB.
goessnd derives temperature and moisture soundings from GOES sounder brightness temperatures datasets (gvarin). The GOES sounder includes 19 channels extending from about 1 to 15 micrometers (Menzel and Purdom, 1994). The algorithm is based on Hayden (1988) and mimics the operational product used by the National Meteorological Center (NMC). The algorithm is a simultaneous physical retrieval of both the temperature and moisture profiles of the atmosphere. It relies heavily on first guess profiles and computes perturbations to these profiles that are consistent with the observed brightness temperatures. The first guess information typically comes from numerical weather analyses (i.e., nowcasts, or initial forecast fields), such as those available from most operational weather centers, including the United States National Weather Service.
The algorithm is designed to derive profiles over a sounding area that typically includes 5 scan lines and 5 samples (see line_sub, samp_sub, and average). The clear areas are distinguished from the cloudy areas based on the cloud screening criteria discussed in Section 2.c of Hayden (1988). If the number of clear scenes is sufficient (see min_clear), then the average brightness temperature for each channel over all the clear scenes in the sounding area is determined. From this mean clear-sky sounding, the temperature and moisture profiles are determined from an iterated inversion of the radiative transfer equation (see max_pass, gamma, wv_scale). This inversion is designed to produce temperature and moisture perturbations to the first guess profiles that bring the new profiles more in line with the mean clear-sky brightness temperatures.
Due to the heavy reliance on the first guess information, it is suggested that the most current guess or forecast information be used in the retrieval process. The present release of goessnd was tested using first-guess analysis data from NMC's global data assimilation system (gdas). These data can be obtained via anonymous ftp from nic.fb4.noaa.gov. The specific files used with goessnd were obtained from the pub/fnl directory. Examples of the two files needed for first guess information are gdas2.T12Z.PGrbF00 and gdas1.T00Z.sstgrb. The first file is a zero-hour forecast (i.e. analysis or nowcast; F00) for time 12Z (GMT). The "2" indicates a spatial resolution of 2.5 degrees and the "P" indicates the analysis is interpolated to a pressure vertical coordinate system. The "Grb" indicates the data are in the WMO Gridded Binary Format (GRIB; Stackpole, 1994). This type of GRIB file contains several three-dimensional fields and several horizontal two-dimensional fields that describe the basic state of the weather forecasting model at its initial time step. This initial state was determined from a spatial interpolation of the most recent weather observations to the current time (e.g. 12Z) using the model physics as constraints for the interpolation. The second file is also a GRIB file which contains the sea surface temperature. In this case, it pertains to time 00Z and the "1" indicates a spatial resolution of 1.0 degree. A shell script (nmcgdas2tdf) is provided to extract the necessary information from files such as these and write it to a TeraScan dataset that can be used by goessnd.
In the cases when/where current weather analysis is not available, a twelve month climatology (ecmwfclim.tdf) is also provided that includes the necessary information for goessnd to initiate a first guess. This information is taken from five years of European Centre for Medium Range Forecasting (EMCWF) analyses. First guess information can also be provided by a user-supplied TeraScan dataset containing the following variable names and units:
Variable Units Dimensions/Description ------------------------------------------------------------------ temperature kelvin 3-D relative_humidity percent 3-D temperature_pressure mb 1-D array of pressure levels relative_humidity_pressure mb 1-D array of pressure levels sea_surface_temperature kelvin 2-D (bad values for land) surface_pressure Pa 2-D surface_temperature kelvin 2-D ~10-50m temperature surface_relative_humidity percent 2-D ~10-50m relative humidity
The horizontal coverage of the first guess dataset must be sufficient to cover the region of the GOES sounder pass being processed, and a valid earth transform must exist for the above variables. The number of pressure levels for the 3-D variables and their exact pressure values are arbitrary, however typically these should extend from 100 mb or above, down to 1000 mb. For pressure levels above that supplied by the first guess data, a simple climatology is blended into the first guess temperature and relative humidity data. Note the last dimension of all the variables above should be the sample dimension (e.g. longitude), the second to last dimension should be the line dimension (e.g. latitude), and, in the case of the 3-D variables, the third to last dimension should be the vertical dimension (e.g. pressure). If goessnd is to assume the supplied first guess dataset represents a climtology (see climges), then each of the above variables must have an additional dimension (their first dimension) that has length twelve, which represents the months of the year beginning with January.
The output variables and parameters of the retrieval algorithm include the following:
Variable Units Description ------------------------------------------------------------------ pressure millibars 1-D Pressure Array temperature kelvin Retrieved Temperatures water_vapor gm/kg Retrieved H20 Mixing Ratio ozone ppm-volume Climatological Ozone Used 1stges_temperature kelvin 1st Guess Temperature 1stges_water_vapor gm/kg 1st Guess Water Vapor 1stges_surf_temperature kelvin 1st Guess Surface Temp. 1stges_surf_dewpoint kelvin 1st Guess Surface Dew Pt. elevation meters Elevation geography 0=Ocean;1=Land;3=Lake;4=Marsh skin_temperature kelvin Retrieved Skin Temperature surface_temperature kelvin Retrieved Surf. Temperature surface_dewpoint kelvin surface_pressure millibars cloud_flag cloud_fraction emissivity avg_goes_bt kelvin 1st_ges_bt kelvin tau optional (see parameter weight) phi optional (see parameter weight) alpha optional (see parameter weight) xit optional (see parameter weight)
The number of lines to subsample between soundings or to include in the processing of a single sounding (see average).
The valid range is [1 to 25]. The default value is 5.
The number of samples to subsample between soundings or to include in the processing of a single sounding (see average).
The valid range is [1 to 25]. The default value is 5.
If average=yes, then each sounding is determined from a rectangle of scenes made up of line_sub lines and samp_sub samples. If average=no, then each sounding is determined from a a single scene and line_sub and samp_sub samples are used to subsample the input dataset.
Valid responses are [yes or no]. The default value is yes.
The minimum fraction of clear pixels required to derive a sounding (see average).
The valid range is [1/(line_sub*samp_sub) to 1.0]. The default value is 0.3.
The perturbations to the first guess profiles are produced by an inversion of the radiative transfer equations. This process can be iterated to improve the retrievals on successive iterations.
The valid range is [1 to 10]. The default value is 3
Specifies whether the first guess information comes from a climatology or from recent forecast/analysis data.
Valid responses are [yes or no]. The default value is no.
Specifies to output the Tau, phi, Phi and inverse matrices (Eqn. 2-10; Hayden, 1988) for the last sounding retrieved. This is typically for debugging purposes when running the function for a single retrieval (e.g. 5x5 box) to examine its sensitivity to the input parameters.
Valid responses are [yes or no]. The default value is no.
Specifies the debug level. Increasing the debug level increases the level of debug information. This is primarily for internal use only but user's may find the information helpful.
The valid range is [0 to 4 ]. The default value is 0.
The algorithm estimates the first guess skin temperature using one of several approaches.
Value Method for estimate Channels 0 shortwave day or night depending on solar zenith angle 1 shortwave day (17,18) 2 shortwave night (16,17) 3 longwave + shortwave (17,7,8) 4 longwave split window a (7,8) 5 longwave split window b (6,8)Valid responses are [0 to 5]. The default value is 3.
Name of TeraScan dataset containing the first guess information (see description above).
If climatology=yes, the default value is ecmwfclim.tdf. If climatology=no, there is no default.
Specifies the value of the stability parameter (Eqn. 10, Hayden, 1988). The larger the value, the smaller the perturbations to the first guess profiles. The smaller the value, the more freedom the retrieval has in modifying the first guess profiles.
The valid range is [0.01 to 100.0]. The default value is 50.
Specifies the scaling value for the water vapor in the Phi matrix (See page 702, column 2, Hayden, 1988).
The valid range is [1.0e-8 to 1.0e8]. The default value is 1000.
OPTIONAL. Value used for checking the temperature perturbation produced by the retrieval. If the difference exceeds this specified value, then the retrieval is assumed to have failed.
The valid range is [0.0 to 100.0]. The default value is 50.
OPTIONAL. Value used for checking the absolute difference between the observed brightness temperatures and those produced from the retrieved profiles. If the difference exceeds this specified value, then the retrieval is assumed to have failed.
The valid range is [0.0 to 100.0]. The default value is 50.
OPTIONAL. Maximum sea surface temperature gradient (difference) allowed over the sounding area (see line_sub, samp_sub) box.
The valid range is [0.0 to 20.0]. The default value is 1.
OPTIONAL. Initial/default value for surface emissivity.
The valid range is [0.0 to1.0]. The default value is 0.99.
OPTIONAL. Minimum surface emissivity accepted.
The valid range is [0.0 to 1.0]. The default value is 0.95.
OPTIONAL. Minimum value for error weight (see Eqn. 11; Hayden, 1988).
The valid range is [0.0 to 10000.0]. The default value is 0.15.
OPTIONAL. Factor by which to reduce gamma on each successive iteration. This allows for decreasing the rigidity of the retrieval on each iteration.
The valid range is [1.0 to 100.0]. The default value is 1.
OPTIONAL. A temperature check value between the first guess and retrieval.
The valid range is [0.0 to 1000.0]. The default value is 4.
OPTIONAL. Skin temperature (kelvin) difference check.
The valid range is [0.0 to 15.0]. The default value is 6.
OPTIONAL. Minimum value for for quantity dU (Eqn. 7; Hayden, 1988) to still allow differential to be calculated.
The valid range is [0.0, 10.0]. The default value is 1.0e-5.
OPTIONAL. Maximum value for visible channel (19) to still be considered a clear scene.
The valid range is [0.0 to 100.0]. The default value is 25.
OPTIONAL. Factor that multiplies the current values of U (Eqns. 3,7) in order to derive an expression for dTau/dU [e.g., dU = U*(1+WFAC) - U(1-WFAC)].
The valid range is [0.0 to 10.0]. The default value is 0.2.
OPTIONAL. Factor for setting minimum water vapor in adjustment routine after retrieval. This factor multiplies the pressure (mb) at the level being checked. The maximum between this value and 0.05 is then used as a relative humidity to multiply the saturation mixing ratio. This latter value is used as the minimum mixing ratio.
The valid range is [0.0 to 100.0]. The default value is 0.0001.
OPTIONAL. Specifies whether or not to perform artificial adjustment to water vapor after retrieval to remove values above 95% relative humidity and keep a minimum value as well (see WLIMX above).
Valid responses are [yes or no]. The default value is yes.
OPTIONAL. Specifies whether or not to increase the uncertainty in the check on window brightness temperatures against the first guess surface temperature in the cloud screening routine.
Valid responses are [yes or no]. The default value is no.
OPTIONAL. If set to less than or equal 99, then for an iteration less than this value (iterations start at 1), reload the first guess temperatures and override the retrieval (e.g. for a value of 1, don't reload temperatures).
The valid range is [0 to 99]. The default value is 1.
OPTIONAL. Specifies whether or not to make an adjustment to the brightness temperatures for daylight hours.
Valid responses are [yes or no]. The default value is yes.
OPTIONAL. Specifies whether or not to make bias adjustments to observed brightness temperatures based on statistical comparisons with observations and retrievals.
Valid responses are [yes or no]. The default value is no.
OPTIONAL. Specifies whether or not to set bottom two temperatures equal (e.g. a simple mixed layer).
Valid responses are [yes or no]. The default value is yes.
OPTIONAL. Specifies whether or not to perform surface temperature gradient check in cloud screening routine. If yes, then don't do the check.
Valid responses are [yes or no]. The default value is no.
OPTIONAL. If set to yes, then don't do surface skin temperature retrieval.
Valid responses are [yes or no]. The default value is no.
OPTIONAL. Determines the value to use for the out of bounds check on the retrieved surface water vapor (i.e., Eqn. 8). If set to no, the observed first guess value is used, otherwise the mean of the observed first guess surface water vapor mixing ratio and the retrieved value at the level just above the lowest level is used.
Valid responses are [yes or no]. The default value is no.
% goessnd pass/g8.95222.0246 g8snd line_sub : int ? [5] samp_sub : int ? [5] average : char( 3) ? [yes] min_clear : real ? [0.3] max_pass : int ? [3] gamma : real ? [50] wv_scale : real ? [1000] 5000 climatology : char( 3) ? [no] guess_file : char(255) ? pass/nmc_95080912.tdf debug : int ? [0] Processing:pass/g8.95222.0246 02:45:13.708 95/08/10: Lines: 224 Samples: 361
$TSCANROOT/refdata/goes/climozon.tdf $TSCANROOT/refdata/goes/climprof.tdf $TSCANROOT/refdata/goes/common.tdf $TSCANROOT/refdata/goes/ecmwfclim.tdf $TSCANROOT/refdata/goes/flemcofx.tdf $TSCANROOT/refdata/goes/geography.tdf $TSCANROOT/refdata/goes/goesrtcx.tdf $TSCANROOT/refdata/goes/refatm.tdf $TSCANROOT/refdata/goes/stdatm.tdf
Hayden, C. M., 1988: GOES-VAS Simultaneous Temperature-Moisture Retrieval Algorithm, J. Appl. Meteor., 27, 705-733. Menzel, W.P. and J.F.W. Purdom, Introducing GOES-I: The First of a New Generation of Geostationary Operational Environmental Satellites. Bull. Amer. Meteor. Soc., 75, 757-781. Stackpole, J. D., 1994: The WMO format for the storage of weather product information and the exchange of weather product messages in gridded binary form. Office Note 388. National Meteorological Center, National Oceanic and Atmospheric Administration. (see pub/doc directory on nic.fb4.noaa.gov anonymous ftp).
Last Update: $Date: 2001/06/19 19:27:16 $